Broadcom just printed a quarter that changes the AVGO debate. Per Broadcom's Q1 FY2026 release, AI revenue came in at $8.4B, up 106% year over year, and OpenAI joined the named custom-silicon customer list.
The thesis is no longer a hope story. Management is projecting over $100B AI chip revenue by 2027 with a customer pipeline that can actually deliver it.
So the retail question shifts from if to how. Do you add AVGO now, dollar-cost average over the next two months, or wait for a pullback?
The Q1 Print That Changed the Story
Total Q1 FY2026 revenue was $19.31B, up 29% year over year. The mix is what matters for the AI thesis.
Semiconductor Solutions delivered $12.5B, up 52% YoY, and now represents 65% of total revenue. Infrastructure Software added $6.8B, up 1% YoY, with VMware doing the heavy lifting on the software side.
Custom silicon: $8.4B AI revenue
The AI line within Semiconductor Solutions hit $8.4B, up 106% YoY. That is the standout number, and it confirms the XPU custom-silicon thesis Broadcom has been telling for two years.
According to Futurum Group, the doubling in AI revenue was driven by XPU momentum across hyperscaler customers ramping into volume production. This is no longer a single-customer ramp.
VMware: the steady drumbeat
VMware revenue grew 13% YoY with $9.2B of Total Contract Value booked in Q1 alone. That TCV print is the highest in any single quarter since the acquisition.
For a retail investor, VMware is the boring half of the story that funds the AI half. Predictable software cash flow lets management invest in the XPU roadmap without stretching the balance sheet.
What the 6-Customer Pipeline Actually Means
Broadcom expanded its named custom-silicon customer list from 5 to 6 this quarter. OpenAI is the new addition, alongside Google, Anthropic, Meta, and two unnamed hyperscalers.
Management used this list to anchor a projection of over $100B AI chip revenue by 2027. That is the number every retail AVGO investor should price against.
Anthropic: 1GW today, 3GW by 2027
Anthropic is running roughly 1 gigawatt of TPU compute in 2026. That demand is expected to surge to over 3 gigawatts in 2027.
A 3x scale-up in two years from a single customer is unusual outside of hyperscaler capex cycles. It also explains why Broadcom is comfortable putting a $100B revenue stake in the ground.
OpenAI: first XPU lands in 2027
OpenAI's first-generation XPU with Broadcom deploys in volume in 2027 at over 1 gigawatt of compute. That is a brand-new revenue stream that does not yet show up in the run rate.
Google adds the third leg. 7th-generation Ironwood TPU demand is strong through 2026, and next-generation TPUs are expected to drive an even larger order book in 2027. For context on how the broader hyperscaler stack sits behind this print, see MSFT and GOOGL repositioning on the mega-cap AI side.
Three Ways to Add AVGO Now
AVGO is up about 23% year to date in 2026 and has more than doubled in the last 12 months. That run-up shapes how a retail investor should enter.
The first path is a lump-sum add today. You are buying confirmed momentum into a customer pipeline that has visibility through 2027. The risk is timing into a stock that has already run hard.
The second path is dollar-cost averaging over the next 8 weeks. Split your intended position into 8 weekly tranches, buy on the same day each week, and ignore the price noise. This is the path that respects how extended the chart is without sitting out the thesis.
The third path is to wait for an 8-12% pullback before adding. AVGO has had two such pullbacks in the last 12 months, and both resolved higher within a few weeks. The risk here is that the pullback does not arrive and you miss the next leg.
None of these paths is wrong. The wrong move is going to full size on day one with no plan for what happens if the stock dips 15% on a macro day.
Risks That Could Slow the Thesis
Customer concentration is the largest single risk. Four named hyperscalers drive most of the XPU revenue, and any one of them slowing capex would dent the trajectory.
ASIC demand is also cyclical. If general-purpose GPUs from NVIDIA regain workload share at the expense of custom silicon, the 2027 mix shifts unfavorably. The AMD vs NVDA comparison matters here because both are the substitutes Broadcom's XPU has to keep displacing.
Foundry capacity is a second-order risk. Broadcom's XPU roadmap depends on Taiwan Semiconductor advanced-node capacity, which is already tight through 2027.
Conclusion
The Q1 FY2026 print confirmed a customer-pipeline story rather than a hope story. AI revenue up 106% with OpenAI now named is a different setup from what AVGO looked like even six months ago.
For a retail investor, the decision is no longer if AVGO belongs in the portfolio. It is how to size the entry into a stock that has already run 23% YTD and more than doubled in twelve months.
Open the Gotrade Global app and pick your path. Trade US stocks from $1 with fractional shares so you can lump-sum, dollar-cost average, or stage a pullback entry at exactly the position size your portfolio can handle.
FAQ
How big is the AI revenue line for Broadcom now?
$8.4B in Q1 FY2026, up 106% year over year, sitting inside a $12.5B Semiconductor Solutions segment.
Who are the 6 named custom-silicon customers?
Google, Anthropic, Meta, OpenAI, and two unnamed hyperscalers, with OpenAI added this quarter as the 6th name.
What is Broadcom's 2027 AI revenue projection?
Management is guiding to over $100B AI chip revenue by 2027, anchored to Google Ironwood, Anthropic scaling to 3GW, and OpenAI's first XPU at 1GW.
Should I wait for a pullback to buy AVGO?
An 8-12% pullback would offer a cleaner entry, but a DCA plan over 8 weeks avoids the risk of missing the move if no pullback arrives.
